How to use the AI Readiness Assessment agent skill

What an agent skill is and what this one does

An agent skill is a folder of instructions and supporting files that an AI agent loads when a request matches the skill's purpose. You do not run it like a normal program. You ask the agent for something, and it uses the skill to carry out a defined procedure.

When asked to audit a repository's AI readiness, this skill has the agent run Microsoft's AgentRC readiness scanner on the project. The scanner produces structured JSON results. A custom reporting agent, ai-readiness-reporter, then fills in a fixed HTML template and saves the dashboard at reports/index.html.

• The dashboard shows an overall score, maturity level and grade.
• It breaks down nine pillars: eight under Repo Health and one under AI Setup.
• Each pillar gets an AI relevance tag (High, Medium or Low) and a specific recommendation.
• It lists a prioritised remediation plan and the active scoring policy.
• The skill measures only. It does not change the project's code, and it only creates reports/index.html.

Install the acreadiness-assess skill

Download the skill's ZIP file from the page where you found this guide. Then choose one of the two installation routes below.

For Claude Code, unzip the file into ~/.claude/skills/ to make it available in all your projects, or into a project's .claude/skills/ folder to limit it to that project.

In the Claude apps, you can upload the ZIP instead. Open Settings, go to Capabilities, then Skills, and upload it there.

• Personal install: unzip into ~/.claude/skills/
• Project install: unzip into the project's .claude/skills/ folder
• Claude apps: Settings → Capabilities → Skills, then upload the ZIP

Check the requirements before you run it

The scan needs Node.js 20 or later on your PATH. If you are not sure what you have installed, run node --version in a terminal. The scan also fetches the AgentRC package from GitHub, so the machine needs network access to GitHub.

The reporting step needs a custom agent named ai-readiness-reporter to be available in your runtime. An optional policy file or package can be used for organisation-specific scoring, but it is not required.

• Node.js 20 or later on PATH
• npx access to fetch github:microsoft/agentrc from GitHub
• Network access to GitHub
• The ai-readiness-reporter custom agent available in the runtime
• Optional: a policy file or package for org-specific scoring

Run an assessment with a natural request

You do not need to remember a command. Ask in plain language, for example that you want the AI readiness of the repo assessed, audited or scored. The skill is written to trigger on requests to assess, audit or score a repository's AI readiness.

You can also invoke it directly as a slash command. The skill accepts two optional arguments: --policy, which takes a path or package, and --per-area. The agent runs the scan from the repository root with JSON output.

When the scan finishes, the agent tells you where the report is and summarises the maturity level, overall score, the three lowest-scoring pillars and the single highest-leverage next action. That next action is usually to run the acreadiness-generate-instructions skill.

• Plain request: 'Audit the AI readiness of this repo.'
• Slash command: /acreadiness-assess
• With a policy: /acreadiness-assess --policy ./policies/strict.json
• Open the result: reports/index.html, which works under file://

Use a policy for consistent scoring across repositories

A policy changes how the assessment is scored. Organisations can use one shared policy across several repositories so that results are comparable. If you do not supply one, the agent checks agentrc.config.json for a policies array. If neither is present, the scan uses built-in defaults.

The dashboard shows the active policy, including any disabled or overridden criteria and thresholds. For background on policies, the agent can suggest the acreadiness-policy skill.

• Passed with --policy: that source is used.
• Otherwise: a policies array in agentrc.config.json is used.
• Otherwise: built-in defaults apply.

Use the assessment in CI

The skill's notes recommend the AgentRC command with a fail level for CI gating. The level is a number from 1 to 5. Use this if you want a build to fail when readiness falls below the level you choose.

The skill itself is an interactive assessment. The CI command below is the AgentRC CLI, as named in the skill notes.